Pancreatic cancer causes glucose intolerance and diabetes in up to 80% of patients. Pancreatic cancer induced diabetes (PaCDM) is often asymptomatic, of short duration (<3 years), and occurs at a resectable stage of the cancer. Up to 1% of newly diagnosed diabetics, equal to or >50 years of age may have PaCDM. We propose to establish whether newly elevated fasting blood glucose (FBG) is indicative of underlying pancreatic cancer as evidenced by diagnosis of pancreatic cancer within 3 years of the FBG measurement. If so, this would provide an entirely novel approach to screening for sporadic pancreatic cancer. We hypothesize that the 3-year likelihood of diagnosis of pancreatic cancer will be high in subjects: a) equal to or > 50 years of age with newly elevated FBG, b) with elevated fasting glucose who have known risk factors for pancreatic cancer (e.g. smoking), and/or c) who manifest an abrupt increase in FBG in serial measurements over time. We will identify visits to Mayo Clinic between years 1988 to 2002 by subjects equal to or >50 years of age who resided in its surrounding catchment area and had a routine physical examination that included a FBG. Preliminary data reveal that about 150,000 different subjects made >300,000 such visits during this time period providing >1 million person-years of follow-up. We will electronically retrieve clinical, and laboratory data and examine 3-year follow-up from date of FBG measurement to identify those diagnosed with pancreatic cancer. We expect 340 to 510 pancreatic cancer events in this cohort. Our Specific Aims are: Aim 1A). To test if FBG drawn during routine physical examination predicts likelihood of underlying pancreatic cancer and to test for non-linearity of this association. Aim 1B) To establish a FBG threshold that predicts a high 3-year likelihood of diagnosis of pancreatic cancer. Aim 2). To estimate the extent to which known risk factors for pancreatic cancer (age, smoking, obesity and family history) modify the likelihood of underlying pancreatic cancer in subjects with FBG greater than the threshold defined by Aim 1B. Aim 3). To test if patterns of change in serial measurements of FBG over time predict likelihood of underlying pancreatic cancer. The clinical and research implications of this study are considerable. These data may lead to delineation of individuals who have a high likelihood of existing pancreatic cancer, and who may be ideal candidates for more intensive screening or early detection regimens. The research described in this application is 100% relevant to pancreatic cancer.